Predicting Prostate Cancer Aggressiveness

描述（由申请人提供）：癌症是一种复杂的疾病，由肿瘤细胞、基质细胞和微环境之间的相互作用驱动。我们假设肿瘤的不同细胞成分和每个细胞内的分子信号网络之间的相互作用可以描述侵袭性前列腺癌。我们选择了代表人类前列腺癌基本过程的细胞内和细胞外途径和细胞类型。我们将使用一大批前列腺癌患者的数据。这些输入，使用最先进的方法导出，并在每个细胞的基础上提供，将用于导出多尺度数学模型。这种丰富的人类数据以前是无法实现的，将用于参数化（在多个尺度上）和验证模型。数学建模将生成一个网络三胞胎库（即肿瘤上皮、正常基质和反应性基质的细胞内信号网络），它们的相互作用可以描述患者的预后。网络将使用遗传算法来识别和修复最适合的三胞胎。然后，三胞胎将根据其在生物学相关时期内反映侵袭性或非侵袭性疾病的能力进行选择，并根据其组织化学特征的表现进行分类。最具代表性的三胞胎将通过体内实验对生物学终点进行验证。验证阶段将概括体内数学模型的关键要素，以确定那些与人类疾病功能最相关的模型。这些将在体内进行测试，以做出预测，以证实或反驳这些结果。最强大的模型（那些通过测试和验证阶段的模型）将与人类临床样本的测试队列进行比较，以将其特征与生存终点相关联。我们独特的资源组合和团队专业知识代表了一个无与伦比的环境，提供了一种超越当前应用科学方法限制的协同方法来了解前列腺癌。我们的模型以人类数据开始和结束，确保最终产品将提供对人类前列腺癌的新理解。将解决三个具体目标：具体目标1)开发和参数化前列腺癌的多尺度数学模型具体目标2)候选数学结果的生物学验证和测试。具体目标3)数学结果的临床验证。相关性（见说明）：数学模型有可能作为有用的预后工具，但尚未很好地发展用于癌症进展的研究。通过基于临床样本反褶积显微镜检查的数据丰富输出的模型，将创建和测试具有无与伦比细节的新模型。这将允许开发新的预后工具和治疗策略来控制疾病进展。

DESCRIPTION (provided by applicant): Cancer is a complex disease that is driven by interactions between tumor cells but also stromal cells and the microenvironment. We hypothesize that the interaction between the different cellular components of the tumor and the molecular signaling networks within each cell can delineate aggressive prostate cancer. We have selected intracellular and extracellular pathways and cell types that are representative of fundamental processes in human prostate cancer. We will use data from a large cohort of prostate cancer patients. These inputs, derived using state of the art methodology, and provided on a cell-per-cell basis will be used to derive a multi-scale mathematical model. This wealth of human data has not previously been achievable and will serve to both parameterize (on multiple scales) and validate the model. Mathematical modeling will generate a library of network triplets (i.e. intracellular signaling networks for tumor epithelium, normal stroma and reactive stroma) whose interactions can describe patient outcome. Networks will be selected using a genetic algorithm to identify and fix the fittest triplets. Triplets will then be selected based upon their ability to reflect invasive or non-invasive disease over a biologically relevant time period and subject to triage based upon their representation of histochemical characteristics. The most representative triplets will be validated against biological endpoints using in vivo experimentation. The validation phase will recapitulate key elements of the mathematical model in vivo to identify those models most functionally-relevant to human disease. These will be tested in vivo to make predictions that confirm or refute these results in silico. The most robust models (those that pass the testing and validation phases) will be compared to a test cohort of human clinical samples to correlate their characteristics against survival endpoints. Our unique combination of resources and team expertise represents an unparalleled environment providing a synergistic approach to understand prostate cancer beyond the limitations of currently applied scientific methodology. Our models begin and end with human data, assuring that the final products will provide new understanding of human prostate cancer. Three specific aims will be addressed: Specific aim 1) Develop and Parameterize a Multi-scale Mathematical Model of Prostate Cancer Specific Aim 2) Biological Validation and Testing of Candidate Mathematical Outcomes. Specific Aim 3) Clinical Validation of Mathematical Outcomes. RELEVANCE (See instructions): Mathematical models have the potential to act as useful prognostic tools but have not yet been well developed for the study of cancer progression. By basing models on data-rich outputs from deconvolution microscopy examination of clinical samples new models with unparalleled detail will be created and tested. This will allow for the development of new prognostic tools and therapeutic strategies to control disease progression.